Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
oncallthrow
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
61.
▲
by
oncallthrow
9mo ago
When I ask Claude to do something, it independently, without me even asking or instructing it to, searches the codebase to understand what the convention is. I’ve even found it searching node_modules to find the API of non-public libraries.
62.
▲
by
oncallthrow
9mo ago
In my experience these models (including opus) aren’t very good at “improving” existing code. I’m not exactly sure why, because the code they produce themselves is generally excellent.
63.
▲
by
oncallthrow
9mo ago
Yeah Opus 4.5 is a massive step change in my experience. I feel like I’m working with a peer, not a junior I’m having to direct. I can give it highly ambiguous and poorly specified tasks and it… just does it. I will note that my experience
64.
▲
by
oncallthrow
9mo ago
This is interestingly somewhere where Go really shines, in my experience. Go has no requirement to wrap (or, indeed, even handle at all) errors; yet, despite this, Go codebases I've worked in almost always perform error handling proper
65.
▲
by
oncallthrow
9mo ago
> I don't think that's different though? It is different. You suggested that rakyll told Claude to simply implement a solution that her team already put the legwork into designing. I'm saying that it sounds like Claude pro
66.
▲
by
oncallthrow
9mo ago
I find Claude (and specifically Opus 4.5) insanely impressive and frankly scary, but I think there is some Pareto principle that people ignore when discussing it. Specifically I find it is great at the first 80%, but I struggle getting the
67.
▲
by
oncallthrow
9mo ago
My reading is precisely the opposite, ie that she fed it the problem without the solution and it independently came to the solution they spent months arguing over.
68.
▲
by
oncallthrow
9mo ago
> As a person living in the Americas… I’m surprised at how good this outcome is? Did we just remove a terrible regime in a comparably bloodless way? It's way too early to tell this. I mean, hopefully yes, but it's way, way too
69.
▲
by
oncallthrow
9mo ago
You might think such a video is not a fit for Hacker News. I implore you, go and watch the video, and I think you will agree with me that it is perfect for this site. Enjoy.
70.
▲
$6 Michelin Stock in 60 Minutes [video]
(youtube.com)
5 points
by
oncallthrow
9mo ago
|
3 comments
71.
▲
by
oncallthrow
9mo ago
I think it’s a mix of human and LLM writing
72.
▲
by
oncallthrow
9mo ago
No, Java (or frankly anything JVM-based) on the terminal is a terrible experience.
73.
▲
Nick Kyrgios defeats Anna Sabalenka to win 'Battle of the Sexes' tennis match
(news.sky.com)
4 points
by
oncallthrow
9mo ago
|
0 comments
74.
▲
by
oncallthrow
10mo ago
> A library is something you call. > A framework is some kind of application scaffolding that normally calls you. There is no real distinction between these two.
75.
▲
by
oncallthrow
10mo ago
By the way, the reason behind all of this, like so many ills of our industry, is the completely broken promotion culture.
76.
▲
by
oncallthrow
10mo ago
This. This. This. I currently work in a codebase where so much code has been abstracted away for “cleanliness” that it’s impossible to understand what code is actually running. The worst is when three lines of completely standard code (imme
77.
▲
by
oncallthrow
10mo ago
I was expecting this to be stupid but it’s genuinely funny. I guess LLMs are better at humor than I remember
78.
▲
by
oncallthrow
10mo ago
> most web developers are well aware of the risky parts of the language In my experience this really isn’t true. Most web developers I know are not familiar (enough) with prototype pollution. By the way, this isn’t because they are “dumb
79.
▲
by
oncallthrow
10mo ago
No, we just found Nicola sturgeon’s hacker news account
80.
▲
George Osborne Joins OpenAI
(bbc.co.uk)
9 points
by
oncallthrow
10mo ago
|
1 comments
81.
▲
by
oncallthrow
10mo ago
It sounds to me like you’re just arguing for premature optimization of another kind (specifically, prematurely changing your entire architecture for edge cases that probably won’t ever happen to you).
82.
▲
by
oncallthrow
10mo ago
If you 300ms latency then yes, you defeat this detection mechanism.
83.
▲
by
oncallthrow
10mo ago
Meh, this is a defence in depth measure anyway Edit: also, the supported platforms are ARM and x86. If your code isn’t running on one of those platforms, you probably know what you’re doing.
84.
▲
by
oncallthrow
10mo ago
Wrong, collective action doesn’t change anything. It simply interchanges who gets poorer.
85.
▲
by
oncallthrow
10mo ago
Sorry, which grass roots group exactly campaigned for this? Which party’s manifesto was it on?
86.
▲
by
oncallthrow
10mo ago
Is tens of thousands vast enough for you?
87.
▲
by
oncallthrow
10mo ago
This is IPFS
88.
▲
by
oncallthrow
10mo ago
Yes, currently they are both smarter and dumber than humans.
89.
▲
by
oncallthrow
10mo ago
I think Perl is still more popular even today than Python as a sysadmin language. Late 2000s it certainly was. Maybe Google was different, but across the industry more widely Python was barely used, Perl was used everywhere.
90.
▲
by
oncallthrow
10mo ago
> also, it seems like everything in perl is a string? It's actually somehow even worse than this. The Perl type system is honestly ridiculous. Most veteran Perl developers don't fully understand it. https://blogs.per
More ›